Jonathan Kitchen/DigitalVision via Getty Images
In the 1960s, meteorologist Edward Lorenz was running weather simulations on an early computer system when he realized that a small rounding difference led to extremely divergent weather predictions. He later called this idea the butterfly effect to communicate that small changes in initial conditions, like a butterfly flapping its wings in Nepal, could produce wildly different outcomes, like rain in New York……..Continue reading…
By: Sarah Scoles
Source: Undark
.
Critics:
Weather forecasting or weather prediction is the application of science and technology to predict the conditions of the atmosphere for a given location and time. People have attempted to predict the weather informally for thousands of years and formally since the 19th century.
Weather forecasts are made by collecting quantitative data about the current state of the atmosphere, land, and ocean and using meteorology to project how the atmosphere will change at a given place. Once calculated manually based mainly upon changes in barometric pressure, current weather conditions, and sky conditions or cloud cover, weather forecasting now relies on computer-based models that take many atmospheric factors into account.
Human input is still required to pick the best possible model to base the forecast upon, which involves pattern recognition skills, teleconnections, knowledge of model performance, and knowledge of model biases. The inaccuracy of forecasting is due to the chaotic nature of the atmosphere; the massive computational power required to solve the equations that describe the atmosphere, the land, and the ocean; the error involved in measuring the initial conditions; and an incomplete understanding of atmospheric and related processes.
Hence, forecasts become less accurate as the difference between the current time and the time for which the forecast is being made (the range of the forecast) increases. The use of ensembles and model consensus helps narrow the error and provide confidence in the forecast. There is a vast variety of end uses for weather forecasts. Weather warnings are important because they are used to protect lives and property.
Forecasts based on temperature and precipitation are important to agriculture, and therefore to traders within commodity markets. Temperature forecasts are used by utility companies to estimate demand over coming days. On an everyday basis, many people use weather forecasts to determine what to wear on a given day. Since outdoor activities are severely curtailed by heavy rain, snow and wind chill, forecasts can be used to plan activities around these events, and to plan ahead and survive them.
Weather forecasting is a part of the economy. For example, in 2009, the US spent approximately $5.8 billion on it, producing benefits estimated at six times as much. The simplest method of forecasting the weather, persistence, relies upon today’s conditions to forecast tomorrow’s. This can be valid when the weather achieves a steady state, such as during the summer season in the tropics. This method strongly depends upon the presence of a stagnant weather pattern.
Therefore, when in a fluctuating pattern, it becomes inaccurate. It can be useful in both short- and long-range forecasts. Measurements of barometric pressure and the pressure tendency (the change of pressure over time) have been used in forecasting since the late 19th century. The larger the change in pressure, especially if more than 3.5 hPa (2.6 mmHg), the larger the change in weather can be expected.
If the pressure drop is rapid, a low pressure system is approaching, and there is a greater chance of rain. Rapid pressure rises are associated with improving weather conditions, such as clearing skies. Along with pressure tendency, the condition of the sky is one of the more important parameters used to forecast weather in mountainous areas. Thickening of cloud cover or the invasion of a higher cloud deck is indicative of rain in the near future.
High thin cirrostratus clouds can create halos around the sun or moon, which wind Morning fog portends fair conditions, as rainy conditions are preceded by wind or clouds that prevent fog formation. The approach of a line of thunderstorms could indicate the approach of a cold front. Cloud-free skies are indicative of fair weather for the near future. A bar can indicate a coming tropical cyclone. The use of sky cover in weather prediction has led to various weather lore over the centuries.
The forecasting of the weather for the following six hours is often referred to as nowcasting. In this time range it is possible to forecast smaller features such as individual showers and thunderstorms with reasonable accuracy, as well as other features too small to be resolved by a computer model. A human given the latest radar, satellite and observational data will be able to make a better analysis of the small scale features present and so will be able to make a more accurate forecast for the following few hours.
However, there are now expert systems using those data and mesoscale numerical model to make better extrapolation, including evolution of those features in time. Accuweather is known for a Minute-Cast, which is a minute-by-minute precipitation forecast for the next two hours. In the past, human forecasters were responsible for generating the weather forecast based upon available observations.
Today, human input is generally confined to choosing a model based on various parameters, such as model biases and performance. Using a consensus of forecast models, as well as ensemble members of the various models, can help reduce forecast error. However, regardless how small the average error becomes with any individual system, large errors within any particular piece of guidance are still possible on any given model run.
Humans are required to interpret the model data into weather forecasts that are understandable to the end user. Humans can use knowledge of local effects that may be too small in size to be resolved by the model to add information to the forecast. While increasing accuracy of forecasting models implies that humans may no longer be needed in the forecasting process at some point in the future, there is currently still a need for human intervention.
The analog technique is a complex way of making a forecast, requiring the forecaster to remember a previous weather event that is expected to be mimicked by an upcoming event. What makes it a difficult technique to use is that there is rarely a perfect analog for an event in the future.Some call this type of forecasting pattern recognition. It remains a useful method of observing rainfall over data voids such as oceans,as well as the forecasting of precipitation amounts and distribution in the future.
A similar technique is used in medium range forecasting, which is known as teleconnections, when systems in other locations are used to help pin down the location of another system within the surrounding regime. An example of teleconnections are by using El Niño-Southern Oscillation (ENSO) related phenomena. Initial attempts to use artificial intelligence began in the 2010s.
Huawei’s Pangu-Weather model, Google’s GraphCast, WindBorne’s WeatherMesh model, Nvidia’s FourCastNet, and the European Centre for Medium-Range Weather Forecasts’ Artificial Intelligence/Integrated Forecasting System, or AIFS all appeared in 2022–2023. In 2024, AIFS started to publish real-time forecasts, showing specific skill at predicting hurricane tracks, but lower-performing on the intensity changes of such storms relative to physics-based models.
Such models use no physics-based atmosphere modeling or large language models. Instead, they learn purely from data such as the ECMWF re-analysis ERA5.These models typically require far less compute than physics-based models. Microsoft’s Aurora system offers global 10-day weather and 5-day air pollution (CO2, NO, NO2, SO2, O3, and particulates) forecasts with claimed accuracy similar to physics-based models, but at orders-of-magnitude lower cost.
Aurora was trained on more than a million hours of data from six weather/climate models. In 2024, a group of researchers at Google’s DeepMind AI research laboratories published a paper in Nature to describe their machine-learning model, called GenCast, that is expected to produce more accurate forecasts than the best traditional weather forecasting systems. In a study conducted using the AIFS, Lang et al. (2024) presented 30-day ensemble simulations of the Madden-Julia Oscillation.
Leave a Reply